Foundation Models

OpenAI’s GPT-5.6 family (Sol flagship, Terra balanced, Luna cost-efficient) for complex reasoning, coding, multimodal understanding, and agentic tool use. The gpt-5.6 API alias routes to Sol. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol; GPT-6 Sol and Luna (2026-09-22) are earlier cost tiers. This entry remains GPT-5.6, which is still available.

OpenAI’s GPT-6 generation. Astra (API id gpt-6-astra, 2026-09-03) is the peak for computer use, coding agents, professional artifacts, science, and cybersecurity-adjacent defender workflows. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol for complex coding and professional work at the same $2/$10 list price as GPT-6 Sol, with cached input at $0.10. GPT-6 Sol (gpt-6-sol, 2026-09-22) and Luna (gpt-6-luna) are the earlier cost tiers. ChatGPT Plus gets Astra in Work and Codex; Chat “GPT-6 Pro” is Pro/Business/Enterprise. GPT-6.1 Sol, Sol, and Luna are in the API and in Work and Codex for paid plans. From 2026-10-07, GPT-6 with Intelligent UI (interactive charts, forms, buttons, and small tools inside answers) powers the ChatGPT Chat tab: GPT-6 Sol for Plus, Pro, Business, and Enterprise, and GPT-6 Luna for Free and Go from 2026-10-08. GPT-5.6 remains available. The GPT-6.1 Sol model page does not say gpt-6-sol was removed.

Anthropic’s Claude Opus 5.5 (API id claude-opus-5-5, released 2026-09-22) for complex agentic coding, enterprise work, long-context analysis, and careful instruction following. Anthropic says it matches Fable 5.1 on most work at a lower price. Fable 5.1 remains a separate higher list-price SKU. Sonnet 5.5 (claude-sonnet-5-5, 2026-09-28) is the current Sonnet at the same $2/$10 list price as Sonnet 5. Haiku 5.5 (claude-haiku-5-5, 2026-10-07) is the current Haiku.

Anthropic’s Claude Sonnet 5.5 (API id claude-sonnet-5-5, released 2026-09-28) for well-scoped everyday coding, agents, and documents. List price matches Sonnet 5 at $2/$10 per 1M tokens. Opus 5.5 remains the high-stakes tier; Haiku 5.5 (2026-10-07) is the low-cost subagent and high-volume tier.

Anthropic’s Claude Fable 5.1 — the most capable widely released Claude for long-horizon agents, deep reasoning, and demanding coding workflows. Mythos 5.1 is the same underlying model with relaxed cyber/life-sciences safeguards for trusted-access programs.

Anthropic’s fast, cost-efficient Claude tier for high-volume chat, classification, extraction, and sub-agent steps where latency and price matter more than peak reasoning. Claude Haiku 5.5 (API id claude-haiku-5-5, released 2026-10-07) is the current Haiku: Anthropic’s fastest model at standard speed, the first Haiku with an adjustable effort setting, and around 75% cheaper to run than Haiku 4.5 on average. Sonnet 5.5 and Opus 5.5 remain better for complex agentic coding.

Stable Gemini API id gemini-4-argon, released 13 October 2026. Google’s model page describes it as the flagship for complex software engineering, deep mathematical and scientific problems, autonomous multi-step agents, and long decode. Thinking can be set to low, medium, or high. The long-decode guide names medium as the default for most coding and agentic work. It documents function calling, structured outputs, code execution, search grounding, URL context, and file search. Computer use is in preview. The model page does not document native MCP support. Image generation, audio generation, and the Live API are not supported.

Google’s Pro-class Gemini for hard reasoning and native multimodal work. Prefer API id gemini-3.1-pro-preview; Gemini 3.5 Pro remains partner-testing. This slug is not Gemini 4 Argon (gemini-4-argon, public 2026-10-13). On the Gemini API, gemini-2.5-pro has no announced shutdown date and is not deprecated; access is limited to projects that have already used it.

DeepSeek’s MoE general model — strong open-weight performance on coding and knowledge tasks with competitive API pricing.

DeepSeek’s reasoning-focused model trained with reinforcement learning for multi-step math, science, and coding problem solving.

DeepSeek’s V4 generation — live Flash is V4.1-Flash-1017 (API id deepseek-flash, upgraded 2026-10-17): Causal Encoder–Decoder MoE, 552B backbone (8B active prefill / 16B decode), native image+text, and 1M context. API id and prices are unchanged from the 2026-09-10 Flash list. MIT weights on Hugging Face. Legacy deepseek-v4-flash and deepseek-v4-flash-vision-exp aliases temporarily route here. deepseek-v4-pro still serves V4-Pro-0813 after 2026-09-14 at unchanged Pro rates (DeepSeek withdrew the planned Flash reroute).

Moonshot’s Kimi K3 — 2.8T MoE (104B active) open-weight multimodal agentic model with 1M context, native vision, and strong long-horizon coding. Weights on Hugging Face under the Kimi K3 License.

Meta Superintelligence Labs’ Muse Wonder (API id muse-wonder-1.0, 2026-10-12) — closed model for longer agentic coding and professional tasks on the Meta Model API and in Muse Code. It sits above Muse Spark 1.3, which remains the lower-priced API. Muse Wonder Full is a Muse Code subscriber variant; the public pricing page lists only muse-wonder-1.0. Preview. Closed weights.

Meta Superintelligence Labs’ Muse Spark 1.3 — closed multimodal model for agentic tasks, long-horizon coding, computer use, and 1M-context workflows via Muse Code and the Meta Model API. As of 2026-10-12 it is the lower-priced API under Muse Wonder (muse-wonder-1.0), not Meta’s peak. Succeeds Spark 1.2; max reasoning is still withheld for safety testing. Muse Spark Open (2026-10-15) is a separate weight release, not this API id.

Meta Superintelligence Labs’ Muse Glimmer — Apache-2.0 ~30B dense multimodal agent model for on-device and single-GPU local agents. Sibling to closed Muse Spark and Muse Wonder; distinct from Llama 4.

Alibaba’s Qwen3 family spanning Qwen3.8-Max (2.4T MoE / 95B active; live cloud alias qwen3.8-max routes to the 0902 snapshot as of 2026-09-05), open Qwen3.8-27B (dense VLM, Apache-2.0), Qwen3.8-Flash-Next (125B / 6B active multimodal MoE + 51B n-gram embeddings)—a Qwen4 architecture preview—and Qwen3.8-Omni-Flash (hosted 2026-09-17: text/image/audio/video in, text out, 1M context; API id qwen3.8-omni-flash). Production Qwen3.8-Flash on QwenCloud adds 1M-default context and built-in tools atop the Flash-Next design.

Meta’s Llama 4 family — open-weight multimodal models designed for research and commercial use under Meta’s community license.

Sarvam’s flagship open-weight reasoning model. A Mixture-of-Experts transformer with 10.3B active parameters, trained from scratch in India, and used to power the Indus assistant.

Sarvam’s efficient open-weight reasoning MoE. About 2.4B active parameters, built for real-time conversational agents on the Samvaad platform and for local or GPU-constrained deployment.

OpenAI’s GPT-5.6 family (Sol flagship, Terra balanced, Luna cost-efficient) for complex reasoning, coding, multimodal understanding, and agentic tool use. The gpt-5.6 API alias routes to Sol. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol; GPT-6 Sol and Luna (2026-09-22) are earlier cost tiers. This entry remains GPT-5.6, which is still available.

OpenAI’s GPT-6 generation. Astra (API id gpt-6-astra, 2026-09-03) is the peak for computer use, coding agents, professional artifacts, science, and cybersecurity-adjacent defender workflows. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol for complex coding and professional work at the same $2/$10 list price as GPT-6 Sol, with cached input at $0.10. GPT-6 Sol (gpt-6-sol, 2026-09-22) and Luna (gpt-6-luna) are the earlier cost tiers. ChatGPT Plus gets Astra in Work and Codex; Chat “GPT-6 Pro” is Pro/Business/Enterprise. GPT-6.1 Sol, Sol, and Luna are in the API and in Work and Codex for paid plans. From 2026-10-07, GPT-6 with Intelligent UI (interactive charts, forms, buttons, and small tools inside answers) powers the ChatGPT Chat tab: GPT-6 Sol for Plus, Pro, Business, and Enterprise, and GPT-6 Luna for Free and Go from 2026-10-08. GPT-5.6 remains available. The GPT-6.1 Sol model page does not say gpt-6-sol was removed.

Anthropic’s Claude Opus 5.5 (API id claude-opus-5-5, released 2026-09-22) for complex agentic coding, enterprise work, long-context analysis, and careful instruction following. Anthropic says it matches Fable 5.1 on most work at a lower price. Fable 5.1 remains a separate higher list-price SKU. Sonnet 5.5 (claude-sonnet-5-5, 2026-09-28) is the current Sonnet at the same $2/$10 list price as Sonnet 5. Haiku 5.5 (claude-haiku-5-5, 2026-10-07) is the current Haiku.

Anthropic’s Claude Sonnet 5.5 (API id claude-sonnet-5-5, released 2026-09-28) for well-scoped everyday coding, agents, and documents. List price matches Sonnet 5 at $2/$10 per 1M tokens. Opus 5.5 remains the high-stakes tier; Haiku 5.5 (2026-10-07) is the low-cost subagent and high-volume tier.

Anthropic’s Claude Fable 5.1 — the most capable widely released Claude for long-horizon agents, deep reasoning, and demanding coding workflows. Mythos 5.1 is the same underlying model with relaxed cyber/life-sciences safeguards for trusted-access programs.

Anthropic’s fast, cost-efficient Claude tier for high-volume chat, classification, extraction, and sub-agent steps where latency and price matter more than peak reasoning. Claude Haiku 5.5 (API id claude-haiku-5-5, released 2026-10-07) is the current Haiku: Anthropic’s fastest model at standard speed, the first Haiku with an adjustable effort setting, and around 75% cheaper to run than Haiku 4.5 on average. Sonnet 5.5 and Opus 5.5 remain better for complex agentic coding.

Stable Gemini API id gemini-4-argon, released 13 October 2026. Google’s model page describes it as the flagship for complex software engineering, deep mathematical and scientific problems, autonomous multi-step agents, and long decode. Thinking can be set to low, medium, or high. The long-decode guide names medium as the default for most coding and agentic work. It documents function calling, structured outputs, code execution, search grounding, URL context, and file search. Computer use is in preview. The model page does not document native MCP support. Image generation, audio generation, and the Live API are not supported.

Google’s Pro-class Gemini for hard reasoning and native multimodal work. Prefer API id gemini-3.1-pro-preview; Gemini 3.5 Pro remains partner-testing. This slug is not Gemini 4 Argon (gemini-4-argon, public 2026-10-13). On the Gemini API, gemini-2.5-pro has no announced shutdown date and is not deprecated; access is limited to projects that have already used it.

Google’s Gemini 3.8 Flash workhorse — fast, token-efficient multimodal model for agentic workflows, coding, and high-throughput apps where latency and cost matter. Succeeds 3.7 Flash (GA Aug 2026); 3.7 remains supported for efficiency-first workloads.

Meta Superintelligence Labs’ Muse Wonder (API id muse-wonder-1.0, 2026-10-12) — closed model for longer agentic coding and professional tasks on the Meta Model API and in Muse Code. It sits above Muse Spark 1.3, which remains the lower-priced API. Muse Wonder Full is a Muse Code subscriber variant; the public pricing page lists only muse-wonder-1.0. Preview. Closed weights.

Meta Superintelligence Labs’ Muse Spark 1.3 — closed multimodal model for agentic tasks, long-horizon coding, computer use, and 1M-context workflows via Muse Code and the Meta Model API. As of 2026-10-12 it is the lower-priced API under Muse Wonder (muse-wonder-1.0), not Meta’s peak. Succeeds Spark 1.2; max reasoning is still withheld for safety testing. Muse Spark Open (2026-10-15) is a separate weight release, not this API id.

Mistral’s flagship large model for enterprise reasoning, multilingual chat, and function calling via La Plateforme and cloud partners. Large 3 is the generally available line; Mistral Large 4 (`mistral-large-4`) entered public preview on 2026-10-06 as a natively multimodal MoE with 1.05T total / 52B active parameters, a 1.6B vision encoder, and a 1M-token context, with open weights promised by the end of October.

xAI’s Grok 4.7 — frontier coding and long-running agentic model (API id grok-4.7) with 500K context, vision, and strong tool use. Available via the xAI API, Grok Build, Cursor, GitHub Copilot, and partners such as OpenRouter.

Cohere’s Command R+ model optimized for retrieval-augmented generation, enterprise search, and multilingual business assistants.

DeepSeek’s MoE general model — strong open-weight performance on coding and knowledge tasks with competitive API pricing.

DeepSeek’s reasoning-focused model trained with reinforcement learning for multi-step math, science, and coding problem solving.

DeepSeek’s V4 generation — live Flash is V4.1-Flash-1017 (API id deepseek-flash, upgraded 2026-10-17): Causal Encoder–Decoder MoE, 552B backbone (8B active prefill / 16B decode), native image+text, and 1M context. API id and prices are unchanged from the 2026-09-10 Flash list. MIT weights on Hugging Face. Legacy deepseek-v4-flash and deepseek-v4-flash-vision-exp aliases temporarily route here. deepseek-v4-pro still serves V4-Pro-0813 after 2026-09-14 at unchanged Pro rates (DeepSeek withdrew the planned Flash reroute).

Moonshot’s Kimi K3 — 2.8T MoE (104B active) open-weight multimodal agentic model with 1M context, native vision, and strong long-horizon coding. Weights on Hugging Face under the Kimi K3 License.

Meta Superintelligence Labs’ Muse Glimmer — Apache-2.0 ~30B dense multimodal agent model for on-device and single-GPU local agents. Sibling to closed Muse Spark and Muse Wonder; distinct from Llama 4.

Alibaba’s Qwen3 family spanning Qwen3.8-Max (2.4T MoE / 95B active; live cloud alias qwen3.8-max routes to the 0902 snapshot as of 2026-09-05), open Qwen3.8-27B (dense VLM, Apache-2.0), Qwen3.8-Flash-Next (125B / 6B active multimodal MoE + 51B n-gram embeddings)—a Qwen4 architecture preview—and Qwen3.8-Omni-Flash (hosted 2026-09-17: text/image/audio/video in, text out, 1M context; API id qwen3.8-omni-flash). Production Qwen3.8-Flash on QwenCloud adds 1M-default context and built-in tools atop the Flash-Next design.

Meta’s Llama 4 family — open-weight multimodal models designed for research and commercial use under Meta’s community license.

Mistral’s sparse Mixture-of-Experts open models (e.g. Mixtral 8x7B / 8x22B) — efficient high-quality text generation for self-hosting.

Sarvam’s flagship open-weight reasoning model. A Mixture-of-Experts transformer with 10.3B active parameters, trained from scratch in India, and used to power the Indus assistant.

Sarvam’s efficient open-weight reasoning MoE. About 2.4B active parameters, built for real-time conversational agents on the Samvaad platform and for local or GPU-constrained deployment.

Sarvam’s earlier multilingual model: a 24B hybrid-reasoning post-train of Mistral Small, released under Apache-2.0 in May 2025. API id sarvam-m.

Microsoft’s Phi family of small language models — high capability per parameter for on-device, edge, and cost-sensitive deployments.

Xiaomi’s MiMo-V2.6 family — MIT-licensed native-omnimodal MoE with 1M context. Flagship open checkpoint is MiMo-V2.6-Pro-RL (1.02T total / 42B activated); Flash-RL is 309B / 15B activated. Hosted API ids: mimo-v2.6-pro, mimo-v2.6-flash, mimo-v2.6-pro-ultraspeed. Xiaomi publishes an Artificial Analysis Intelligence Index score of 46 for Pro; that figure is vendor-published and is not a DataAIHub ranking.

Reflection’s Beam 1.0 — Apache-2.0 text-only sparse MoE (501B total, 23B active, 1M context) for agentic coding and reasoning. Weights shipped 2026-10-16 with day-0 support in Hugging Face Transformers, vLLM, and SGLang. This is not the beam.cloud GPU tool.

OpenAI’s GPT-5.6 family (Sol flagship, Terra balanced, Luna cost-efficient) for complex reasoning, coding, multimodal understanding, and agentic tool use. The gpt-5.6 API alias routes to Sol. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol; GPT-6 Sol and Luna (2026-09-22) are earlier cost tiers. This entry remains GPT-5.6, which is still available.

OpenAI’s GPT-6 generation. Astra (API id gpt-6-astra, 2026-09-03) is the peak for computer use, coding agents, professional artifacts, science, and cybersecurity-adjacent defender workflows. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol for complex coding and professional work at the same $2/$10 list price as GPT-6 Sol, with cached input at $0.10. GPT-6 Sol (gpt-6-sol, 2026-09-22) and Luna (gpt-6-luna) are the earlier cost tiers. ChatGPT Plus gets Astra in Work and Codex; Chat “GPT-6 Pro” is Pro/Business/Enterprise. GPT-6.1 Sol, Sol, and Luna are in the API and in Work and Codex for paid plans. From 2026-10-07, GPT-6 with Intelligent UI (interactive charts, forms, buttons, and small tools inside answers) powers the ChatGPT Chat tab: GPT-6 Sol for Plus, Pro, Business, and Enterprise, and GPT-6 Luna for Free and Go from 2026-10-08. GPT-5.6 remains available. The GPT-6.1 Sol model page does not say gpt-6-sol was removed.

Anthropic’s Claude Opus 5.5 (API id claude-opus-5-5, released 2026-09-22) for complex agentic coding, enterprise work, long-context analysis, and careful instruction following. Anthropic says it matches Fable 5.1 on most work at a lower price. Fable 5.1 remains a separate higher list-price SKU. Sonnet 5.5 (claude-sonnet-5-5, 2026-09-28) is the current Sonnet at the same $2/$10 list price as Sonnet 5. Haiku 5.5 (claude-haiku-5-5, 2026-10-07) is the current Haiku.

Anthropic’s Claude Sonnet 5.5 (API id claude-sonnet-5-5, released 2026-09-28) for well-scoped everyday coding, agents, and documents. List price matches Sonnet 5 at $2/$10 per 1M tokens. Opus 5.5 remains the high-stakes tier; Haiku 5.5 (2026-10-07) is the low-cost subagent and high-volume tier.

Anthropic’s Claude Fable 5.1 — the most capable widely released Claude for long-horizon agents, deep reasoning, and demanding coding workflows. Mythos 5.1 is the same underlying model with relaxed cyber/life-sciences safeguards for trusted-access programs.

Stable Gemini API id gemini-4-argon, released 13 October 2026. Google’s model page describes it as the flagship for complex software engineering, deep mathematical and scientific problems, autonomous multi-step agents, and long decode. Thinking can be set to low, medium, or high. The long-decode guide names medium as the default for most coding and agentic work. It documents function calling, structured outputs, code execution, search grounding, URL context, and file search. Computer use is in preview. The model page does not document native MCP support. Image generation, audio generation, and the Live API are not supported.

Google’s Pro-class Gemini for hard reasoning and native multimodal work. Prefer API id gemini-3.1-pro-preview; Gemini 3.5 Pro remains partner-testing. This slug is not Gemini 4 Argon (gemini-4-argon, public 2026-10-13). On the Gemini API, gemini-2.5-pro has no announced shutdown date and is not deprecated; access is limited to projects that have already used it.

DeepSeek’s MoE general model — strong open-weight performance on coding and knowledge tasks with competitive API pricing.

DeepSeek’s reasoning-focused model trained with reinforcement learning for multi-step math, science, and coding problem solving.

DeepSeek’s V4 generation — live Flash is V4.1-Flash-1017 (API id deepseek-flash, upgraded 2026-10-17): Causal Encoder–Decoder MoE, 552B backbone (8B active prefill / 16B decode), native image+text, and 1M context. API id and prices are unchanged from the 2026-09-10 Flash list. MIT weights on Hugging Face. Legacy deepseek-v4-flash and deepseek-v4-flash-vision-exp aliases temporarily route here. deepseek-v4-pro still serves V4-Pro-0813 after 2026-09-14 at unchanged Pro rates (DeepSeek withdrew the planned Flash reroute).

Moonshot’s Kimi K3 — 2.8T MoE (104B active) open-weight multimodal agentic model with 1M context, native vision, and strong long-horizon coding. Weights on Hugging Face under the Kimi K3 License.

Meta Superintelligence Labs’ Muse Wonder (API id muse-wonder-1.0, 2026-10-12) — closed model for longer agentic coding and professional tasks on the Meta Model API and in Muse Code. It sits above Muse Spark 1.3, which remains the lower-priced API. Muse Wonder Full is a Muse Code subscriber variant; the public pricing page lists only muse-wonder-1.0. Preview. Closed weights.

Meta Superintelligence Labs’ Muse Spark 1.3 — closed multimodal model for agentic tasks, long-horizon coding, computer use, and 1M-context workflows via Muse Code and the Meta Model API. As of 2026-10-12 it is the lower-priced API under Muse Wonder (muse-wonder-1.0), not Meta’s peak. Succeeds Spark 1.2; max reasoning is still withheld for safety testing. Muse Spark Open (2026-10-15) is a separate weight release, not this API id.

Meta Superintelligence Labs’ Muse Glimmer — Apache-2.0 ~30B dense multimodal agent model for on-device and single-GPU local agents. Sibling to closed Muse Spark and Muse Wonder; distinct from Llama 4.

Alibaba’s Qwen3 family spanning Qwen3.8-Max (2.4T MoE / 95B active; live cloud alias qwen3.8-max routes to the 0902 snapshot as of 2026-09-05), open Qwen3.8-27B (dense VLM, Apache-2.0), Qwen3.8-Flash-Next (125B / 6B active multimodal MoE + 51B n-gram embeddings)—a Qwen4 architecture preview—and Qwen3.8-Omni-Flash (hosted 2026-09-17: text/image/audio/video in, text out, 1M context; API id qwen3.8-omni-flash). Production Qwen3.8-Flash on QwenCloud adds 1M-default context and built-in tools atop the Flash-Next design.

Meta’s Llama 4 family — open-weight multimodal models designed for research and commercial use under Meta’s community license.

Mistral’s flagship large model for enterprise reasoning, multilingual chat, and function calling via La Plateforme and cloud partners. Large 3 is the generally available line; Mistral Large 4 (`mistral-large-4`) entered public preview on 2026-10-06 as a natively multimodal MoE with 1.05T total / 52B active parameters, a 1.6B vision encoder, and a 1M-token context, with open weights promised by the end of October.

Sarvam’s flagship open-weight reasoning model. A Mixture-of-Experts transformer with 10.3B active parameters, trained from scratch in India, and used to power the Indus assistant.

Sarvam’s efficient open-weight reasoning MoE. About 2.4B active parameters, built for real-time conversational agents on the Samvaad platform and for local or GPU-constrained deployment.

Sarvam’s earlier multilingual model: a 24B hybrid-reasoning post-train of Mistral Small, released under Apache-2.0 in May 2025. API id sarvam-m.

xAI’s Grok 4.7 — frontier coding and long-running agentic model (API id grok-4.7) with 500K context, vision, and strong tool use. Available via the xAI API, Grok Build, Cursor, GitHub Copilot, and partners such as OpenRouter.

Cohere’s Command R+ model optimized for retrieval-augmented generation, enterprise search, and multilingual business assistants.

Microsoft’s Phi family of small language models — high capability per parameter for on-device, edge, and cost-sensitive deployments.

Xiaomi’s MiMo-V2.6 family — MIT-licensed native-omnimodal MoE with 1M context. Flagship open checkpoint is MiMo-V2.6-Pro-RL (1.02T total / 42B activated); Flash-RL is 309B / 15B activated. Hosted API ids: mimo-v2.6-pro, mimo-v2.6-flash, mimo-v2.6-pro-ultraspeed. Xiaomi publishes an Artificial Analysis Intelligence Index score of 46 for Pro; that figure is vendor-published and is not a DataAIHub ranking.

Reflection’s Beam 1.0 — Apache-2.0 text-only sparse MoE (501B total, 23B active, 1M context) for agentic coding and reasoning. Weights shipped 2026-10-16 with day-0 support in Hugging Face Transformers, vLLM, and SGLang. This is not the beam.cloud GPU tool.

OpenAI’s GPT-5.6 family (Sol flagship, Terra balanced, Luna cost-efficient) for complex reasoning, coding, multimodal understanding, and agentic tool use. The gpt-5.6 API alias routes to Sol. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol; GPT-6 Sol and Luna (2026-09-22) are earlier cost tiers. This entry remains GPT-5.6, which is still available.

OpenAI’s GPT-6 generation. Astra (API id gpt-6-astra, 2026-09-03) is the peak for computer use, coding agents, professional artifacts, science, and cybersecurity-adjacent defender workflows. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol for complex coding and professional work at the same $2/$10 list price as GPT-6 Sol, with cached input at $0.10. GPT-6 Sol (gpt-6-sol, 2026-09-22) and Luna (gpt-6-luna) are the earlier cost tiers. ChatGPT Plus gets Astra in Work and Codex; Chat “GPT-6 Pro” is Pro/Business/Enterprise. GPT-6.1 Sol, Sol, and Luna are in the API and in Work and Codex for paid plans. From 2026-10-07, GPT-6 with Intelligent UI (interactive charts, forms, buttons, and small tools inside answers) powers the ChatGPT Chat tab: GPT-6 Sol for Plus, Pro, Business, and Enterprise, and GPT-6 Luna for Free and Go from 2026-10-08. GPT-5.6 remains available. The GPT-6.1 Sol model page does not say gpt-6-sol was removed.

Anthropic’s Claude Opus 5.5 (API id claude-opus-5-5, released 2026-09-22) for complex agentic coding, enterprise work, long-context analysis, and careful instruction following. Anthropic says it matches Fable 5.1 on most work at a lower price. Fable 5.1 remains a separate higher list-price SKU. Sonnet 5.5 (claude-sonnet-5-5, 2026-09-28) is the current Sonnet at the same $2/$10 list price as Sonnet 5. Haiku 5.5 (claude-haiku-5-5, 2026-10-07) is the current Haiku.

Anthropic’s Claude Sonnet 5.5 (API id claude-sonnet-5-5, released 2026-09-28) for well-scoped everyday coding, agents, and documents. List price matches Sonnet 5 at $2/$10 per 1M tokens. Opus 5.5 remains the high-stakes tier; Haiku 5.5 (2026-10-07) is the low-cost subagent and high-volume tier.

Anthropic’s Claude Fable 5.1 — the most capable widely released Claude for long-horizon agents, deep reasoning, and demanding coding workflows. Mythos 5.1 is the same underlying model with relaxed cyber/life-sciences safeguards for trusted-access programs.

Anthropic’s fast, cost-efficient Claude tier for high-volume chat, classification, extraction, and sub-agent steps where latency and price matter more than peak reasoning. Claude Haiku 5.5 (API id claude-haiku-5-5, released 2026-10-07) is the current Haiku: Anthropic’s fastest model at standard speed, the first Haiku with an adjustable effort setting, and around 75% cheaper to run than Haiku 4.5 on average. Sonnet 5.5 and Opus 5.5 remain better for complex agentic coding.

Stable Gemini API id gemini-4-argon, released 13 October 2026. Google’s model page describes it as the flagship for complex software engineering, deep mathematical and scientific problems, autonomous multi-step agents, and long decode. Thinking can be set to low, medium, or high. The long-decode guide names medium as the default for most coding and agentic work. It documents function calling, structured outputs, code execution, search grounding, URL context, and file search. Computer use is in preview. The model page does not document native MCP support. Image generation, audio generation, and the Live API are not supported.

Google’s Pro-class Gemini for hard reasoning and native multimodal work. Prefer API id gemini-3.1-pro-preview; Gemini 3.5 Pro remains partner-testing. This slug is not Gemini 4 Argon (gemini-4-argon, public 2026-10-13). On the Gemini API, gemini-2.5-pro has no announced shutdown date and is not deprecated; access is limited to projects that have already used it.

Google’s Gemini 3.8 Flash workhorse — fast, token-efficient multimodal model for agentic workflows, coding, and high-throughput apps where latency and cost matter. Succeeds 3.7 Flash (GA Aug 2026); 3.7 remains supported for efficiency-first workloads.

DeepSeek’s MoE general model — strong open-weight performance on coding and knowledge tasks with competitive API pricing.

DeepSeek’s reasoning-focused model trained with reinforcement learning for multi-step math, science, and coding problem solving.

DeepSeek’s V4 generation — live Flash is V4.1-Flash-1017 (API id deepseek-flash, upgraded 2026-10-17): Causal Encoder–Decoder MoE, 552B backbone (8B active prefill / 16B decode), native image+text, and 1M context. API id and prices are unchanged from the 2026-09-10 Flash list. MIT weights on Hugging Face. Legacy deepseek-v4-flash and deepseek-v4-flash-vision-exp aliases temporarily route here. deepseek-v4-pro still serves V4-Pro-0813 after 2026-09-14 at unchanged Pro rates (DeepSeek withdrew the planned Flash reroute).

Moonshot’s Kimi K3 — 2.8T MoE (104B active) open-weight multimodal agentic model with 1M context, native vision, and strong long-horizon coding. Weights on Hugging Face under the Kimi K3 License.

Meta Superintelligence Labs’ Muse Wonder (API id muse-wonder-1.0, 2026-10-12) — closed model for longer agentic coding and professional tasks on the Meta Model API and in Muse Code. It sits above Muse Spark 1.3, which remains the lower-priced API. Muse Wonder Full is a Muse Code subscriber variant; the public pricing page lists only muse-wonder-1.0. Preview. Closed weights.

Meta Superintelligence Labs’ Muse Spark 1.3 — closed multimodal model for agentic tasks, long-horizon coding, computer use, and 1M-context workflows via Muse Code and the Meta Model API. As of 2026-10-12 it is the lower-priced API under Muse Wonder (muse-wonder-1.0), not Meta’s peak. Succeeds Spark 1.2; max reasoning is still withheld for safety testing. Muse Spark Open (2026-10-15) is a separate weight release, not this API id.

Meta Superintelligence Labs’ Muse Glimmer — Apache-2.0 ~30B dense multimodal agent model for on-device and single-GPU local agents. Sibling to closed Muse Spark and Muse Wonder; distinct from Llama 4.

Alibaba’s Qwen3 family spanning Qwen3.8-Max (2.4T MoE / 95B active; live cloud alias qwen3.8-max routes to the 0902 snapshot as of 2026-09-05), open Qwen3.8-27B (dense VLM, Apache-2.0), Qwen3.8-Flash-Next (125B / 6B active multimodal MoE + 51B n-gram embeddings)—a Qwen4 architecture preview—and Qwen3.8-Omni-Flash (hosted 2026-09-17: text/image/audio/video in, text out, 1M context; API id qwen3.8-omni-flash). Production Qwen3.8-Flash on QwenCloud adds 1M-default context and built-in tools atop the Flash-Next design.

Meta’s Llama 4 family — open-weight multimodal models designed for research and commercial use under Meta’s community license.

Mistral’s flagship large model for enterprise reasoning, multilingual chat, and function calling via La Plateforme and cloud partners. Large 3 is the generally available line; Mistral Large 4 (`mistral-large-4`) entered public preview on 2026-10-06 as a natively multimodal MoE with 1.05T total / 52B active parameters, a 1.6B vision encoder, and a 1M-token context, with open weights promised by the end of October.

Mistral’s sparse Mixture-of-Experts open models (e.g. Mixtral 8x7B / 8x22B) — efficient high-quality text generation for self-hosting.

Sarvam’s flagship open-weight reasoning model. A Mixture-of-Experts transformer with 10.3B active parameters, trained from scratch in India, and used to power the Indus assistant.

Sarvam’s efficient open-weight reasoning MoE. About 2.4B active parameters, built for real-time conversational agents on the Samvaad platform and for local or GPU-constrained deployment.

xAI’s Grok 4.7 — frontier coding and long-running agentic model (API id grok-4.7) with 500K context, vision, and strong tool use. Available via the xAI API, Grok Build, Cursor, GitHub Copilot, and partners such as OpenRouter.

Microsoft’s Phi family of small language models — high capability per parameter for on-device, edge, and cost-sensitive deployments.

Xiaomi’s MiMo-V2.6 family — MIT-licensed native-omnimodal MoE with 1M context. Flagship open checkpoint is MiMo-V2.6-Pro-RL (1.02T total / 42B activated); Flash-RL is 309B / 15B activated. Hosted API ids: mimo-v2.6-pro, mimo-v2.6-flash, mimo-v2.6-pro-ultraspeed. Xiaomi publishes an Artificial Analysis Intelligence Index score of 46 for Pro; that figure is vendor-published and is not a DataAIHub ranking.

Reflection’s Beam 1.0 — Apache-2.0 text-only sparse MoE (501B total, 23B active, 1M context) for agentic coding and reasoning. Weights shipped 2026-10-16 with day-0 support in Hugging Face Transformers, vLLM, and SGLang. This is not the beam.cloud GPU tool.

OpenAI’s GPT-5.6 family (Sol flagship, Terra balanced, Luna cost-efficient) for complex reasoning, coding, multimodal understanding, and agentic tool use. The gpt-5.6 API alias routes to Sol. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol; GPT-6 Sol and Luna (2026-09-22) are earlier cost tiers. This entry remains GPT-5.6, which is still available.

OpenAI’s GPT-6 generation. Astra (API id gpt-6-astra, 2026-09-03) is the peak for computer use, coding agents, professional artifacts, science, and cybersecurity-adjacent defender workflows. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol for complex coding and professional work at the same $2/$10 list price as GPT-6 Sol, with cached input at $0.10. GPT-6 Sol (gpt-6-sol, 2026-09-22) and Luna (gpt-6-luna) are the earlier cost tiers. ChatGPT Plus gets Astra in Work and Codex; Chat “GPT-6 Pro” is Pro/Business/Enterprise. GPT-6.1 Sol, Sol, and Luna are in the API and in Work and Codex for paid plans. From 2026-10-07, GPT-6 with Intelligent UI (interactive charts, forms, buttons, and small tools inside answers) powers the ChatGPT Chat tab: GPT-6 Sol for Plus, Pro, Business, and Enterprise, and GPT-6 Luna for Free and Go from 2026-10-08. GPT-5.6 remains available. The GPT-6.1 Sol model page does not say gpt-6-sol was removed.

Anthropic’s Claude Opus 5.5 (API id claude-opus-5-5, released 2026-09-22) for complex agentic coding, enterprise work, long-context analysis, and careful instruction following. Anthropic says it matches Fable 5.1 on most work at a lower price. Fable 5.1 remains a separate higher list-price SKU. Sonnet 5.5 (claude-sonnet-5-5, 2026-09-28) is the current Sonnet at the same $2/$10 list price as Sonnet 5. Haiku 5.5 (claude-haiku-5-5, 2026-10-07) is the current Haiku.

Anthropic’s Claude Sonnet 5.5 (API id claude-sonnet-5-5, released 2026-09-28) for well-scoped everyday coding, agents, and documents. List price matches Sonnet 5 at $2/$10 per 1M tokens. Opus 5.5 remains the high-stakes tier; Haiku 5.5 (2026-10-07) is the low-cost subagent and high-volume tier.

Anthropic’s Claude Fable 5.1 — the most capable widely released Claude for long-horizon agents, deep reasoning, and demanding coding workflows. Mythos 5.1 is the same underlying model with relaxed cyber/life-sciences safeguards for trusted-access programs.

Anthropic’s fast, cost-efficient Claude tier for high-volume chat, classification, extraction, and sub-agent steps where latency and price matter more than peak reasoning. Claude Haiku 5.5 (API id claude-haiku-5-5, released 2026-10-07) is the current Haiku: Anthropic’s fastest model at standard speed, the first Haiku with an adjustable effort setting, and around 75% cheaper to run than Haiku 4.5 on average. Sonnet 5.5 and Opus 5.5 remain better for complex agentic coding.

Stable Gemini API id gemini-4-argon, released 13 October 2026. Google’s model page describes it as the flagship for complex software engineering, deep mathematical and scientific problems, autonomous multi-step agents, and long decode. Thinking can be set to low, medium, or high. The long-decode guide names medium as the default for most coding and agentic work. It documents function calling, structured outputs, code execution, search grounding, URL context, and file search. Computer use is in preview. The model page does not document native MCP support. Image generation, audio generation, and the Live API are not supported.

Google’s Pro-class Gemini for hard reasoning and native multimodal work. Prefer API id gemini-3.1-pro-preview; Gemini 3.5 Pro remains partner-testing. This slug is not Gemini 4 Argon (gemini-4-argon, public 2026-10-13). On the Gemini API, gemini-2.5-pro has no announced shutdown date and is not deprecated; access is limited to projects that have already used it.

Google’s Gemini 3.8 Flash workhorse — fast, token-efficient multimodal model for agentic workflows, coding, and high-throughput apps where latency and cost matter. Succeeds 3.7 Flash (GA Aug 2026); 3.7 remains supported for efficiency-first workloads.

DeepSeek’s V4 generation — live Flash is V4.1-Flash-1017 (API id deepseek-flash, upgraded 2026-10-17): Causal Encoder–Decoder MoE, 552B backbone (8B active prefill / 16B decode), native image+text, and 1M context. API id and prices are unchanged from the 2026-09-10 Flash list. MIT weights on Hugging Face. Legacy deepseek-v4-flash and deepseek-v4-flash-vision-exp aliases temporarily route here. deepseek-v4-pro still serves V4-Pro-0813 after 2026-09-14 at unchanged Pro rates (DeepSeek withdrew the planned Flash reroute).

Moonshot’s Kimi K3 — 2.8T MoE (104B active) open-weight multimodal agentic model with 1M context, native vision, and strong long-horizon coding. Weights on Hugging Face under the Kimi K3 License.

Meta Superintelligence Labs’ Muse Wonder (API id muse-wonder-1.0, 2026-10-12) — closed model for longer agentic coding and professional tasks on the Meta Model API and in Muse Code. It sits above Muse Spark 1.3, which remains the lower-priced API. Muse Wonder Full is a Muse Code subscriber variant; the public pricing page lists only muse-wonder-1.0. Preview. Closed weights.

Meta Superintelligence Labs’ Muse Spark 1.3 — closed multimodal model for agentic tasks, long-horizon coding, computer use, and 1M-context workflows via Muse Code and the Meta Model API. As of 2026-10-12 it is the lower-priced API under Muse Wonder (muse-wonder-1.0), not Meta’s peak. Succeeds Spark 1.2; max reasoning is still withheld for safety testing. Muse Spark Open (2026-10-15) is a separate weight release, not this API id.

Meta Superintelligence Labs’ Muse Glimmer — Apache-2.0 ~30B dense multimodal agent model for on-device and single-GPU local agents. Sibling to closed Muse Spark and Muse Wonder; distinct from Llama 4.

Alibaba’s Qwen3 family spanning Qwen3.8-Max (2.4T MoE / 95B active; live cloud alias qwen3.8-max routes to the 0902 snapshot as of 2026-09-05), open Qwen3.8-27B (dense VLM, Apache-2.0), Qwen3.8-Flash-Next (125B / 6B active multimodal MoE + 51B n-gram embeddings)—a Qwen4 architecture preview—and Qwen3.8-Omni-Flash (hosted 2026-09-17: text/image/audio/video in, text out, 1M context; API id qwen3.8-omni-flash). Production Qwen3.8-Flash on QwenCloud adds 1M-default context and built-in tools atop the Flash-Next design.

Meta’s Llama 4 family — open-weight multimodal models designed for research and commercial use under Meta’s community license.

Xiaomi’s MiMo-V2.6 family — MIT-licensed native-omnimodal MoE with 1M context. Flagship open checkpoint is MiMo-V2.6-Pro-RL (1.02T total / 42B activated); Flash-RL is 309B / 15B activated. Hosted API ids: mimo-v2.6-pro, mimo-v2.6-flash, mimo-v2.6-pro-ultraspeed. Xiaomi publishes an Artificial Analysis Intelligence Index score of 46 for Pro; that figure is vendor-published and is not a DataAIHub ranking.

Stable Gemini API id gemini-4-argon, released 13 October 2026. Google’s model page describes it as the flagship for complex software engineering, deep mathematical and scientific problems, autonomous multi-step agents, and long decode. Thinking can be set to low, medium, or high. The long-decode guide names medium as the default for most coding and agentic work. It documents function calling, structured outputs, code execution, search grounding, URL context, and file search. Computer use is in preview. The model page does not document native MCP support. Image generation, audio generation, and the Live API are not supported.

Google’s Pro-class Gemini for hard reasoning and native multimodal work. Prefer API id gemini-3.1-pro-preview; Gemini 3.5 Pro remains partner-testing. This slug is not Gemini 4 Argon (gemini-4-argon, public 2026-10-13). On the Gemini API, gemini-2.5-pro has no announced shutdown date and is not deprecated; access is limited to projects that have already used it.

Xiaomi’s MiMo-V2.6 family — MIT-licensed native-omnimodal MoE with 1M context. Flagship open checkpoint is MiMo-V2.6-Pro-RL (1.02T total / 42B activated); Flash-RL is 309B / 15B activated. Hosted API ids: mimo-v2.6-pro, mimo-v2.6-flash, mimo-v2.6-pro-ultraspeed. Xiaomi publishes an Artificial Analysis Intelligence Index score of 46 for Pro; that figure is vendor-published and is not a DataAIHub ranking.

OpenAI’s GPT-5.6 family (Sol flagship, Terra balanced, Luna cost-efficient) for complex reasoning, coding, multimodal understanding, and agentic tool use. The gpt-5.6 API alias routes to Sol. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol; GPT-6 Sol and Luna (2026-09-22) are earlier cost tiers. This entry remains GPT-5.6, which is still available.

OpenAI’s GPT-6 generation. Astra (API id gpt-6-astra, 2026-09-03) is the peak for computer use, coding agents, professional artifacts, science, and cybersecurity-adjacent defender workflows. GPT-6.1 Sol (gpt-6.1-sol, 2026-09-29) is the current lower-cost Sol for complex coding and professional work at the same $2/$10 list price as GPT-6 Sol, with cached input at $0.10. GPT-6 Sol (gpt-6-sol, 2026-09-22) and Luna (gpt-6-luna) are the earlier cost tiers. ChatGPT Plus gets Astra in Work and Codex; Chat “GPT-6 Pro” is Pro/Business/Enterprise. GPT-6.1 Sol, Sol, and Luna are in the API and in Work and Codex for paid plans. From 2026-10-07, GPT-6 with Intelligent UI (interactive charts, forms, buttons, and small tools inside answers) powers the ChatGPT Chat tab: GPT-6 Sol for Plus, Pro, Business, and Enterprise, and GPT-6 Luna for Free and Go from 2026-10-08. GPT-5.6 remains available. The GPT-6.1 Sol model page does not say gpt-6-sol was removed.

Anthropic’s Claude Opus 5.5 (API id claude-opus-5-5, released 2026-09-22) for complex agentic coding, enterprise work, long-context analysis, and careful instruction following. Anthropic says it matches Fable 5.1 on most work at a lower price. Fable 5.1 remains a separate higher list-price SKU. Sonnet 5.5 (claude-sonnet-5-5, 2026-09-28) is the current Sonnet at the same $2/$10 list price as Sonnet 5. Haiku 5.5 (claude-haiku-5-5, 2026-10-07) is the current Haiku.

Anthropic’s Claude Sonnet 5.5 (API id claude-sonnet-5-5, released 2026-09-28) for well-scoped everyday coding, agents, and documents. List price matches Sonnet 5 at $2/$10 per 1M tokens. Opus 5.5 remains the high-stakes tier; Haiku 5.5 (2026-10-07) is the low-cost subagent and high-volume tier.

Anthropic’s Claude Fable 5.1 — the most capable widely released Claude for long-horizon agents, deep reasoning, and demanding coding workflows. Mythos 5.1 is the same underlying model with relaxed cyber/life-sciences safeguards for trusted-access programs.

Anthropic’s fast, cost-efficient Claude tier for high-volume chat, classification, extraction, and sub-agent steps where latency and price matter more than peak reasoning. Claude Haiku 5.5 (API id claude-haiku-5-5, released 2026-10-07) is the current Haiku: Anthropic’s fastest model at standard speed, the first Haiku with an adjustable effort setting, and around 75% cheaper to run than Haiku 4.5 on average. Sonnet 5.5 and Opus 5.5 remain better for complex agentic coding.

Stable Gemini API id gemini-4-argon, released 13 October 2026. Google’s model page describes it as the flagship for complex software engineering, deep mathematical and scientific problems, autonomous multi-step agents, and long decode. Thinking can be set to low, medium, or high. The long-decode guide names medium as the default for most coding and agentic work. It documents function calling, structured outputs, code execution, search grounding, URL context, and file search. Computer use is in preview. The model page does not document native MCP support. Image generation, audio generation, and the Live API are not supported.

Google’s Pro-class Gemini for hard reasoning and native multimodal work. Prefer API id gemini-3.1-pro-preview; Gemini 3.5 Pro remains partner-testing. This slug is not Gemini 4 Argon (gemini-4-argon, public 2026-10-13). On the Gemini API, gemini-2.5-pro has no announced shutdown date and is not deprecated; access is limited to projects that have already used it.

Google’s Gemini 3.8 Flash workhorse — fast, token-efficient multimodal model for agentic workflows, coding, and high-throughput apps where latency and cost matter. Succeeds 3.7 Flash (GA Aug 2026); 3.7 remains supported for efficiency-first workloads.

DeepSeek’s V4 generation — live Flash is V4.1-Flash-1017 (API id deepseek-flash, upgraded 2026-10-17): Causal Encoder–Decoder MoE, 552B backbone (8B active prefill / 16B decode), native image+text, and 1M context. API id and prices are unchanged from the 2026-09-10 Flash list. MIT weights on Hugging Face. Legacy deepseek-v4-flash and deepseek-v4-flash-vision-exp aliases temporarily route here. deepseek-v4-pro still serves V4-Pro-0813 after 2026-09-14 at unchanged Pro rates (DeepSeek withdrew the planned Flash reroute).

Moonshot’s Kimi K3 — 2.8T MoE (104B active) open-weight multimodal agentic model with 1M context, native vision, and strong long-horizon coding. Weights on Hugging Face under the Kimi K3 License.

Meta Superintelligence Labs’ Muse Wonder (API id muse-wonder-1.0, 2026-10-12) — closed model for longer agentic coding and professional tasks on the Meta Model API and in Muse Code. It sits above Muse Spark 1.3, which remains the lower-priced API. Muse Wonder Full is a Muse Code subscriber variant; the public pricing page lists only muse-wonder-1.0. Preview. Closed weights.

Meta Superintelligence Labs’ Muse Spark 1.3 — closed multimodal model for agentic tasks, long-horizon coding, computer use, and 1M-context workflows via Muse Code and the Meta Model API. As of 2026-10-12 it is the lower-priced API under Muse Wonder (muse-wonder-1.0), not Meta’s peak. Succeeds Spark 1.2; max reasoning is still withheld for safety testing. Muse Spark Open (2026-10-15) is a separate weight release, not this API id.

Meta Superintelligence Labs’ Muse Glimmer — Apache-2.0 ~30B dense multimodal agent model for on-device and single-GPU local agents. Sibling to closed Muse Spark and Muse Wonder; distinct from Llama 4.

Alibaba’s Qwen3 family spanning Qwen3.8-Max (2.4T MoE / 95B active; live cloud alias qwen3.8-max routes to the 0902 snapshot as of 2026-09-05), open Qwen3.8-27B (dense VLM, Apache-2.0), Qwen3.8-Flash-Next (125B / 6B active multimodal MoE + 51B n-gram embeddings)—a Qwen4 architecture preview—and Qwen3.8-Omni-Flash (hosted 2026-09-17: text/image/audio/video in, text out, 1M context; API id qwen3.8-omni-flash). Production Qwen3.8-Flash on QwenCloud adds 1M-default context and built-in tools atop the Flash-Next design.

Meta’s Llama 4 family — open-weight multimodal models designed for research and commercial use under Meta’s community license.

Xiaomi’s MiMo-V2.6 family — MIT-licensed native-omnimodal MoE with 1M context. Flagship open checkpoint is MiMo-V2.6-Pro-RL (1.02T total / 42B activated); Flash-RL is 309B / 15B activated. Hosted API ids: mimo-v2.6-pro, mimo-v2.6-flash, mimo-v2.6-pro-ultraspeed. Xiaomi publishes an Artificial Analysis Intelligence Index score of 46 for Pro; that figure is vendor-published and is not a DataAIHub ranking.

Google’s Gemini 3.8 Flash workhorse — fast, token-efficient multimodal model for agentic workflows, coding, and high-throughput apps where latency and cost matter. Succeeds 3.7 Flash (GA Aug 2026); 3.7 remains supported for efficiency-first workloads.

Meta Superintelligence Labs’ Muse Glimmer — Apache-2.0 ~30B dense multimodal agent model for on-device and single-GPU local agents. Sibling to closed Muse Spark and Muse Wonder; distinct from Llama 4.

Mistral’s sparse Mixture-of-Experts open models (e.g. Mixtral 8x7B / 8x22B) — efficient high-quality text generation for self-hosting.

Sarvam’s efficient open-weight reasoning MoE. About 2.4B active parameters, built for real-time conversational agents on the Samvaad platform and for local or GPU-constrained deployment.

Sarvam’s earlier multilingual model: a 24B hybrid-reasoning post-train of Mistral Small, released under Apache-2.0 in May 2025. API id sarvam-m.

Microsoft’s Phi family of small language models — high capability per parameter for on-device, edge, and cost-sensitive deployments.